Cat Carrier Pet Patreon: Fan Support

Pet carrier production desk · Updated 2026-10-06 · 17 min read

Pet carrier OEM and ODM work on cat carrier platforms splits at the pattern - OEM builds to your drawing, ODM adapts an existing platform and removes the tooling cost.

Tier Structure Read as a Production Specification

A support tier ladder is, in manufacturing terms, a variant matrix with a quantity attached to each cell. Reading it that way turns a merchandising question into a specification, and the specification is what drives cost.

A typical ladder has four tiers. An entry tier of 200-600 backers receiving a small item or nothing at all; a mid tier of 80-250 receiving a decorated unit; an upper tier of 30-120 receiving a premium variant; and a top tier of 10-40 receiving a numbered edition. The quantities are small and they move every cycle as backers upgrade or lapse.

The cost consequence is that each tier wants a different decoration level, and decoration levels do not scale down linearly. A woven label at 0.12-0.35 USD needs 100-500 units to be economical; a moulded patch at 0.55-1.40 plus a 180-450 USD mould needs 300 or more to amortise the tool.

The bottom row is the key to the whole structure. All four tiers share one base unit, so the 500-piece minimum is met once by the aggregate rather than four times by each tier. A programme with 320 backers across all tiers still places a 500-unit base order and holds the surplus as blanks for the next cycle.

The upper tier row is where programmes get into trouble. A moulded patch needs a 180-450 USD mould and a 300-unit minimum, and a 30-120 unit tier amortises neither. The specification answer is to use a stock patch shape with a custom colour, which removes the mould cost, or to move the upper tier onto embroidery at 0.45-1.20.

Cycle-to-cycle volatility is the planning problem. Tier populations move 12-28% between cycles as backers upgrade and lapse, so a base run sized to the current cycle is wrong by the next. Holding blanks sized at 1.3-1.9 times the present cycle absorbs that without a new production run.

Four tiers share one 500-unit base order; only the upper tier at 30-120 units has a method-minimum problem, solved with a stock patch shape or embroidery.

Aggregating Tier Quantities to Clear the Minimum

The minimum of 500 pieces per colourway is met by the aggregate, but the arithmetic still has to be planned, because the surplus does not disappear — it sits as blanks and has to be sized deliberately.

Three aggregation structures are used. Pooling by base, where all tiers share one colourway and one base run. Pooling across cycles, where two or three reward cycles are produced in one run and decorated per cycle. And pooling by shared variant, where two adjacent tiers take the same decoration and are distinguished only by packaging or an insert.

The third is underrated. If the mid and upper tiers take identical decoration and differ only by a numbered card, their quantities combine for decoration purposes and the upper tier's premium feel comes from the insert rather than from a second decoration process.

The bottom row shows what unpooled looks like: sixteen base runs a year at 500 units each, or 8,000 units against an annual demand of 1,280-4,040. It is included only to show why no programme does it.

The two-cycle row is the usual answer. At 900-1,600 units per run with 580-1,180 blanks carried, the programme absorbs 12-28% tier volatility across two cycles and carries 1,531-6,226 USD a year of inventory cost. Against the saving on the volume ladder, that is comfortably worthwhile.

The decision rule on how far to pool is the ratio of carrying cost to volume-tier saving. Pooling from one cycle to two moves the unit cost from 17.20-19.80 to 14.80-16.40, a saving of 2.40-3.40 USD per unit, against carrying cost of 0.22-0.44 USD per unit per month. Two cycles of pooling pays back in roughly seven months of holding.

Pool across two cycles: 900-1,600 units per run, 580-1,180 blanks carried at 1,531-6,226 USD a year, absorbing 12-28% tier volatility.

Numbered Editions and the Cost of Per-Unit Variation

A numbered edition is the standard top-tier reward and it is the most expensive decoration in this set, because the number changes on every unit. That is variable data applied to the product rather than to a card, and the cost structure is completely different.

Embroidering a sequential number requires the machine to stop, load a new digitising file and restart for each unit. That is 40-110 seconds of machine time per unit against 8-20 for a fixed mark, which is where the 0.85-2.40 USD comes from. On a 40-unit top tier that is 34-96 USD; on the same quantity as a card insert it would be 2-11.

The alternatives are a numbered card at 0.06-0.28, a numbered woven label at 0.12-0.35 with a 500-unit minimum, or a hand-applied number tag at 0.35-0.95. All three give the collectable effect at a fraction of the cost, and all three are indistinguishable to the recipient.

The 40-unit column is where a top tier sits, and it shows the numbered card or the transfer as the sensible choice at 14-101 USD for the edition. Sequential embroidery at 94-236 USD for forty units is a premium of 72-187 USD for an effect the recipient can barely distinguish.

Edition size should be fixed before production, not after. A declared edition of 250 that actually ships 180 creates a credibility problem; a declared edition of 180 that ships 180 is clean. The production record should state the number produced and the number destroyed as rejects, and both figures belong in the insert.

Rejects in a numbered edition need a rule. A unit rejected at final inspection is replaced from blank stock and takes the next number, and the rejected number is recorded as destroyed rather than reused. Without that rule a declared edition of 200 can contain two units numbered 137.

Number the edition on a card or a transfer at 14-101 USD for forty units; sequential embroidery costs 94-236 for the same effect and fixes the edition size before production starts.

Changeover Behaviour Across a Four-Variant Ladder

Four variants sounds like four setups, and on a custom-run basis it is four production changeovers at 60-200 USD each plus four short cutting runs. On a pooled base it is one setup and four decoration batches, and the difference is most of the economic case.

The decoration batches are not free, though. Each has a changeover of 0-140 USD depending on method, and a first-piece verification. On a 60-unit upper tier that is 0.00-2.33 USD per unit of changeover alone, which is why the smallest tiers should use the cheapest method available rather than the most prestigious.

Sequencing matters more here than in other channels because the tiers ship together. Decorate in descending quantity so the largest tier is finished first and the small tiers follow; the alternative leaves the biggest batch waiting on the smallest and delays the whole cycle by three to eight days.

The last row is the end state for a programme running four or more cycles a year. With blanks already in the fulfilment country, the cycle collapses to decoration and packing, 6-24 days, and the production changeover disappears entirely. The cost is the carrying cost of the in-country holding at 0.22-0.44 USD per unit per month.

The shared-upper row is the cheap structural fix. If the upper and mid tiers take the same decoration and differ by an insert, one decoration batch is removed, saving 140 USD of changeover and three to four days. It costs nothing in perceived value if the insert is well specified.

Cycle time is the number a programme feels. A 61-88 day custom cycle against a reward promised for a specific month means the order is placed before the tier population is known, which is the root cause of over- and under-buying in this channel. Shortening the cycle to 39-58 days, or to 6-24 with blanks held, lets the order follow the actual count.

A pooled base takes a four-tier cycle from 61-88 days to 39-58, and to 6-24 with blanks held in country; sharing decoration between adjacent tiers saves a further 140 USD and 3-4 days.

Direct-to-Backer Shipping: Pack, Transit and Address Integrity

Membership rewards ship one parcel at a time to individual addresses, often across twenty to forty countries. That is the most demanding fulfilment pattern in this set, and it has three cost drivers: packet rate, damage rate and address integrity.

Packet rate is driven by weight band and destination zone. A packed carrier at 1.1-1.9 kg moves in the 1-2 kg band, and the band price varies 6.80-21.40 USD between a domestic zone and a distant international one. Programmes that offer free shipping to all backers are effectively subsidising 6.80-21.40 USD per unit and should price it explicitly.

Damage in individual parcels runs 0.8-2.4% unverified and 0.3-0.9% with drop verification, and the difference is worse here than elsewhere because a damaged replacement ships as another individual parcel at the full packet rate. A damaged unit costs the unit plus 6.80-21.40 USD of replacement freight.

The consolidated row is the structural saving where a programme has more than about fifteen backers in one country. Consolidating to a regional hub and re-dispatching locally costs 4.90-9.80 USD against 12.60-21.40, a saving of 7.70-11.60 USD per unit, and it requires one customs documentation set rather than fifteen.

Address integrity is the operational risk. Backer-supplied addresses change, contain errors and are sometimes incomplete, and a misdelivery costs 12-28 USD to recover plus a replacement. A validation step against a postal database at 0.04-0.12 USD per address, plus a pre-dispatch confirmation request, cuts the failure rate from 1.4-3.6% to 0.3-0.9%.

Customs documentation is the last element. A commercial invoice with an accurate HS classification, a declared value and country of origin is required on every international parcel; a misclassified parcel is held for 4-21 days and often returned. Preparing the documentation as part of the pack at 0.08-0.22 USD per unit is far cheaper than handling holds.

Consolidate to a regional hub above fifteen backers per country, saving 7.70-11.60 USD per unit, and validate addresses at 0.04-0.12 USD to cut failure from 1.4-3.6% to 0.3-0.9%.

Compliance for Individually Shipped Gift Items

A reward shipped directly to a supporter enters each destination country as an import, and the obligations that apply are those of the destination rather than those of a single warehouse market. That multiplies the documentation question by the number of countries in the backer list.

The baseline set is the same everywhere: country of origin, fibre content, care instruction and a stated maximum pet weight. These sit on the unit and do not change by destination. The variable part is chemical compliance and language.

Chemical obligations differ by jurisdiction. A California-destined parcel raises a Proposition 65 question; an EU-destined parcel raises REACH; a UK-destined parcel raises UK REACH. A programme shipping to forty countries cannot hold forty declarations, and the practical answer is to declare against the strictest applicable standard globally, which costs 220-780 USD once rather than per market.

Language is the recurring cost people forget. An instruction insert in the language of the destination costs 0.10-0.42 USD per unit per language, and a programme shipping to twelve language markets either prints twelve versions or ships one multilingual sheet. The multilingual sheet is cheaper at 0.18-0.62 USD and is the standard answer.

Where the carrier is presented as suitable for air travel, the insert should state collapsed and rigid dimensions and reference published cabin rules rather than claiming approval. Guidance on animal import and health requirements is published by the US Centers for Disease Control and Prevention and by the US Department of Agriculture, and neither certifies equipment.

Textile chemistry is declared against OEKO-TEX criteria and structural testing is referenced to ASTM International methods, both issued from the SGS-verified production base so the importer's file is complete at entry.

Declare to the strictest standard globally at 220-780 USD once, ship a multilingual insert at 0.18-0.62 USD, and never claim approval where a dimensional statement is supportable.

Sampling a Tier Ladder

A four-tier ladder needs four approvals, and the temptation is to sample them in parallel. Parallel sampling on a ladder is slower than sequential, because all four come back with corrections at once and the correction round has to be repeated.

The efficient sequence samples the base once and the tiers afterwards. The base sample covers structure, dimensions, materials and colourway and is approved in 6-10 working days. Each tier then needs only a decoration sample at 15-140 USD, approved in 1-6 days.

The top tier is the exception and it should be sampled first among the tiers, because a numbered edition has the most ways to be wrong: number format, placement, font size and the edition statement in the insert. Sampling it first leaves room to correct before the cycle closes.

The combined set at the end is the control that catches what individual samples miss: whether the four tiers actually look like a ladder. A mid and an upper tier that are too similar undermine the tier structure, and that is only visible when they are side by side.

Total sampling cost is 250-740 USD, or 0.39-2.31 USD per unit across a 320-640 unit cycle. On small tiers that is a meaningful per-unit load, which is another argument for pooling cycles so the sampling is amortised over two or three cycles rather than one.

Retained samples matter more in a repeating programme than anywhere else. One sealed unit per tier per cycle, held 12-24 months, is what settles the question of whether cycle four matched cycle one. At unit cost that is 55-79 USD per cycle and it prevents a dispute that costs far more.

Sample the platform once at 45-120 USD, then tiers at 15-140 each with the numbered top tier first; finish with a combined set at 120-300 USD to verify the ladder reads as a ladder.

Cost Model for a Membership Reward Programme

The channel models as follows. Take four cycles a year, 640 units per cycle across four tiers, one base colourway, sea freight to a fulfilment warehouse in the programme's home market, and individual parcels to backers in six countries.

The unpooled version runs sixteen base orders a year at 500 units each — four tiers times four cycles — which is 8,000 units against 2,560 of demand. It is included for completeness rather than as a real option; landed cost is 34.20-52.60 USD per unit with heavy surplus.

The pooled version runs two base orders of 1,280 units, decorates sixteen batches across four cycles, holds the balance, and consolidates parcels to two regional hubs.

The largest single line is the backer parcel, and the largest single saving is consolidating it. Seven of the 13.75-20.41 USD total comes from moving parcels to regional hubs rather than from anything in the factory, which is a useful reminder that in a direct-shipping channel the fulfilment structure dominates the manufacturing structure.

Surplus is the other difference the table understates. The unpooled model produces 8,000 units against 2,560 of demand, so its true cost includes 5,440 units of dead stock. The pooled model carries 580-1,180 blanks, all of which are fungible into the next cycle.

Commercial terms run as standard: MOQ 500 pieces per colourway on the base run, decoration minimums from one unit for transfer and embroidery and 300 for a moulded patch, prototypes in 6-10 working days for the platform and 1-6 days for decoration, bulk production 35-50 days after sample approval, final random inspection to AQL 2.5, T/T 30/70 and FOB Xiamen. Quality management runs under ISO 9001 with BSCI coverage. Pooling saves 13.75-20.41 USD per unit, and more than half of it comes from consolidating backer parcels rather than from production.